188?. 1 PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. , r >81 



branchial arch, and the relations of the gill-membrane of the respective 

 sides. In their analysis (p. 401) they have represented the relationships 

 and characteristics of the several families in the following manner: 



DD. Suborbital with bony stay. (Cottiform fishes.) 

 v. Head not mailed. 



x. Slit behind fourth gill large; body evenly scaled CMridte 



xx. Slit behind fourth gill small or wanting. 



y. Gill-membranes separate, free from isthmus ; ventral fins normally 



formed, mostly i,5 Scorpainidce. 



yy. Gill-membranes broadly joined, attached to the isthmus or not ; veu 



tral fins variously imperfect Cottidw. 



vv. Head mailed, externally bony. 



z. Ventrals few-rayed, close together ; last gill-slit obsolete. Agonidce. 

 zz. Ventrals i,4, on, 5, usually wide apart ; last gill-slit large. Triglidce. 



Those having a suborbital stay but haviug the " breast with a suck- 

 ing-disk" are divided into two families: 



b. Skin smooth ; vertebrae very numerous Liparididce. 



hi. Skin tubercular; vertebne rather few Cyclopteridce. 



It will thus be seen that the characteristics of the families given by 

 Professors Jordan and Gilbert are not of great importance, and we need 

 not be surprised, therefore, that they considered that " the relations of 

 each, especially of the Scorpamidce, Agonidce, and Liparididce with the 

 Coitidce are so close that it is difficult to draw satisfactory boundary 

 lines." The characteristics assigned by them to the families are, how- 

 ever, co-ordinate with osteological characters of far greater importance 

 which confirm their families so far as they go, but it will become evi- 

 dent hereafter they have not gone far enougli, and the families require 

 to be multiplied. The characters of the additional families to be ad- 

 mitted are of fundamental importance and greater than those assigned 

 by the authors to the families admitted by them. 



OWEN. 



Among the statements relating to the skeleton, one occurs which 

 should not be passed over in silence, and which may be aptly noticed 

 in this place. According to Professor Owen (Anatomy of Vertebrates, 

 v. i, p. Ill), the subtectals or " orbitosphenoids" are "sometimes repre- 

 sented by a descending plate of the frontal, as in the Garpike, or by 

 unossified cartilage, as in mail cheeked fishes. 17 



This statement must surely be the result of some confnsiou of notes 

 or misapplication of the name " mail-cheeked fishes." In all the spe- 

 cies of that series which I have examined (and which must have been 

 many more than observed by Professor Owen), the so-called orbito- 

 sphenoids or subtectals are very distinctly developed, and the modifica- 

 tions of those bones have been found to be very useful in the determi- 

 nation of the relationships of the species, as well as for diagnostic 

 purposes. Whatever may have been the basis of observations, the 

 statement at any rate is altogether too sweeping and vague. 



